A/B test across full stack applications
A/B testing is an experimentation process where two variations of software are tested among a randomized, controlled set of users to ensure statistical rigor.
Each A/B test includes a set of metrics as part of the hypothesis, defining the goals of the experiment. In many cases you may want to affect several metrics while not affecting others, called guardrail metrics.
Turn every feature into an A/B test
Turning every feature you develop into an A/B test. Split ties feature flags to customizable metrics from your operational and customer data pipelines, automatically measuring the impact of every feature on every metric you define.
Split takes a statistically rigorous approach to A/B testing, detecting variation across treatment groups to determine whether an experiment was successful and if you should continue rolling it out.